Using Madjik API for A2A

Enable autonomous agent communication and coordination using Madjik's market intelligence.

Overview

A2A (Agent-to-Agent) protocols allow AI agents to communicate and collaborate. Madjik API provides the shared data layer for crypto-focused agent networks.

A2A Architecture with Madjik

┌─────────────┐     ┌─────────────┐     ┌─────────────┐
│  Agent A    │────▶│  Madjik API │◀────│  Agent B    │
│ (Sentiment) │     │  (Shared    │     │  (Risk)     │
└─────────────┘     │   Truth)    │     └─────────────┘
                    └─────────────┘
                          │
                          ▼
                    ┌─────────────┐
                    │  Agent C    │
                    │ (Execution) │
                    └─────────────┘

Use Cases

1. Multi-Agent Trading System

# Agent 1: Sentiment Analyst
class SentimentAgent:
    def analyze(self):
        sentiment = requests.get(
            "https://api.madjik.io/v1/metrics/me10030",
            headers={"Authorization": "Bearer YOUR_API_KEY"}
        ).json()
        return {"signal": "bullish" if sentiment['value'] > 60 else "bearish"}

# Agent 2: Risk Manager
class RiskAgent:
    def assess(self):
        liquidation = requests.get(
            "https://api.madjik.io/v1/metrics/me10016",
            headers={"Authorization": "Bearer YOUR_API_KEY"}
        ).json()
        return {"risk_level": liquidation['value']}

# Agent 3: Executor
class ExecutionAgent:
    def execute(self, sentiment_signal, risk_level):
        if sentiment_signal == "bullish" and risk_level < 50:
            return self.buy()
        elif risk_level > 80:
            return self.reduce_position()

2. Consensus-Based Decisions

def multi_agent_consensus():
    agents = [
        SentimentAgent(),
        WhaleTrackingAgent(),
        TetherMonitorAgent()
    ]
    
    votes = []
    for agent in agents:
        # All agents use Madjik as shared truth
        signal = agent.analyze()
        votes.append(signal)
    
    # Majority vote
    return max(set(votes), key=votes.count)

3. Agent Communication Protocol

# Standardized message format using Madjik metrics
message = {
    "from": "sentiment_agent",
    "to": "execution_agent",
    "timestamp": "2026-01-29T12:00:00Z",
    "data": {
        "madjik_metrics": {
            "ME10030": 72,  # Social Sentiment
            "ME10014": 0.05  # Funding Rate
        },
        "recommendation": "long",
        "confidence": 0.85
    }
}

Best Practices

1. Shared data source - All agents should use Madjik as single source of truth
2. Timestamped data - Ensure agents reference same time window
3. Conflict resolution - Define rules when agents disagree
4. Audit trail - Log all agent decisions with underlying Madjik data

See Also

• AI Integration Guide
• MCP Integration Guide
• All Metrics